| A new technique for the Bell-427 helicopter model identification from flight data tests is here presented. A helicopter model is identified and validated for 22 flight conditions.;The best final score is obtained with the linear method optimized with neural network theory in order to observe the system outputs from its states. The obtained matrices are interpolated to obtain the model for any flight condition.;Key words: identification, validation, helicopter, FAA.;Two major problems are solved in this thesis. The first problem concerns the state equation. A recurrence relationship and then, an optimization based on neural network theory are performed. A manual and automatic tuning of the initial state conditions is done in order to satisfy the FAA rules. The second problem regards the output equation. A comparison among three methods (classical equations of motion, fuzzy logic and linear methods) is performed in order to find the powerful method. |
